by Vas Mylko, Roman Bilusiak
The newest version 18 of Curiosio is out. We are recreating the mental process of a traveler. With the breakthrough AI we are on the way to support the travel planning workflow very well. It means support for both planning and re-planning, both before and during the trip. It will be a digital concierge that is feeling a curious traveler — feeling you. It will be assisting you on your unique never-ending journey to see all places in the world before your die.
You’ll have to be Signed in to use the new workflow and to get My Trips. We use Google, Twitter, Facebook, GitHub for authorized access. We save only profile identifier (some long machine-friendly code) during OAuth authorization and keep an access token as a web browser cookie. Who doesn’t hate those annoying cookie pop-up forms? So we are not doing it during beta
You could jump right away to the new version at https://curiosio.com [use Chrome browser]. Though we recommend you keep reading about how we designed it, and what is the big picture that we envision and are building for.
Travel Planning Workflow
We are describing two use cases below. One for a traveler who wants inspiration from others before his own creativity kicks in; and one for a supertraveler who knows what she wants out of the box.
Use Case 1: Discovery & Inspiration
I am a New Yorker. I’m not going anywhere during the next few weeks but I love to review others’ travels systematically. It’s like a daily cup of tea or coffee for me. I run some lazy searches on the East Coast to explore what’s possible. I like Signature Trips that Curiosio is recommending in the United States. I’d like to collect the promising trip plans for further review. So I’m Liking the candidates, then I’ll decide which of them could be interesting for my family,
Then, I’m going to My Trips, see the “Liked trips” carousel. The interactive trip plans appear in chronological order there— the latest liked plan will be the first on the left. I’m opening a trip to Cape Cod and start augmenting it via the [Supertrip] button. First off, I’m changing the home point to New York City and setting the car option to My Car.
It’s a good practice to Like the search results. Open the trip plan you like and hit the Like button on it. That plan will appear in My Trips. The most recently liked plan will be the leftmost in the carousel in My Trips.
Use Case 2: Full Workflow
I am planning a road trip with Curiosio. I’ve made some anchor decisions — some subset of who’s going, where, when and why. I’ve got some preferences — budget, duration, key points and places, style. I need to discover my options for driving, lodging, sightseeing and activities within those constraints.
I’m searching multiple times for the different items with different constraints. I’m learning what’s possible, understanding the implications, and continue researching according to Curiosio's recommendations. Once I eventually get some “maybe” alternatives I get them as the complete trip plans — each plan with itinerary, route, overnight stays, breakdown by time and cost.
I am collecting the candidate plans, saving the most promising ones for further analysis, sharing these plans with other participants to get their opinion. Sharing the URLs [to the interactive plans] is a geeky way that already worked for a while. Believe it or not — some people don’t even know how to get an URL from a browser that is hiding it. “Liking” is the first friendlier way to collect the plans, and it’s implemented for the first time in this version of Curiosio. “Saving” is another way, not implemented yet.
Artificial Smartness
The codename for v18 was User Loop. Here we will outline some bits of the technological design behind the scenes. We are building a travel digital concierge. It started from computational intelligence, continues to an answering engine, and eventually will evolve into a personal digital concierge. Conceptually there are two feedback loops to build artificial smartness: People-to-People with AI inside, AI-to-AI with People inside.
Trips Liked by People
In this “classical” loop people educate machines. Travelers know the thing of travel — huge and old industry. There is a tremendous amount of domain knowledge. Machines are learning from people, via Machine Learning. Machine help travelers to discover and plan interesting trips.
Travelers like to plan ahead for security and to deviate from the plan by following their curiosity. It’s true for the planning workflow, and true during the journey. It’s human nature to connect associations during a creative or discovery session, like in The Glass Bead Game.
The journey brings experience and emotions. Though the use of time or money might be not so rational for that experience. More could be experienced for the same time & money. Or money or time could be saved on the same experience.
Trips Liked by Machines
Another analogy of the human vs. machine style is a game of chess. Ever noticed that human games are more likable, and beautiful, resembling performing arts? While computer games are strict and dryasdust but highly efficient? Computers always win over humans. Humans know the thing, Computers know the score.
The second loop is AI-to-AI with People inside. It’s People who are learning, so it’s People Learning and Machine Doing. It’s artificial smartness. AI has an advantage over people to look deeper into the future how things could unfold, and what is possible in the unique set of goals with constraints.
Machine Empathy
We need these two loops for a technological goal to build the foundation of the digital Concierge. Simply speaking, a digital concierge is machine empathy plus productivity (fine-tuning, co-design, booking/rebooking, refunding, etc.) The empathic side is called Affective Computing. More magic by Curiosio will be delivered in the next versions.
Netflix of Trips
My Trips page is going to be a Netflix of Trips. There will be many carousels of related trip plans in each. We will get there incrementally. In this release, we implemented “Liked trips”. In future releases, we will implement “Saved trips”, “Recent trips”. Then, we will add “Top trips for you”, “Quirky trips”, “Unexpected trips”, “Trending trips”, “Secret trips”, and so forth.
As soon as we get enough data we will start showing a second carousel on the Landing page and country pages. It will be a leader board from the trips liked by Signed in travelers and by machines/AI. This carousel will be visible to anonymous site visitors. But the number of trip plans there will be limited.
Signed-in travelers will get a real deal. They will get way more travel content on the My Trips page. Numerous trip plans will be organized in multiple carousels like movies and TV series on Netflix. You will be able to configure your own carousels, store what you want, or define criteria so that Curiosio puts relevant trip plans for you. Request what you need in advance and Curiosio will prepare the trip plans for you by the next day morning.
Some trips or whole carousels by Supertravelers will be on the My Trips page. It will be implemented when a sufficient number of Supertravelers are “traveling” on our platform. A lot of plans here. From the list of carousels with quirky Netflix’ish collections to specific and magical Spotify’ish playlists. The artificial smartness [see the previous section] will make it possible.
Big Picture
There is no special graphical interface in Curiosio. To be precise, there are no decorations — the information is the interface. Everything is information. It means zero learning time to start using the tool. Literally, almost everything in the Curiosio trip plan is clickable & tappable. Everything is linked to the English Wikipedia for rich primary information. Wikipedia is linked to Wikivoyage for the travel angle on the points or places. Every route could be exported to Google Maps. Nothing is vogue but nothing will get outdated during the next 20 years.
Current trip plans from the search results are more friendly to machines than to humans. Machines like them as is. People need stories, travel stories. Like those juicy stories from Lonely Planet or Condé Nast Traveler or National Geographic. Travel storytelling will happen in v19 codenamed Natural Language Generation. Can you imagine Lonely Planet writing, designing, and publishing a unique guidebook only for you, only for your special next trip? Curiosio will do this in a minute. So stay tuned for the v19 news.
The times when travel stories were told or put in writing are gone. Modern travelers want video storytelling. How could new travelers hear about us? Videos from the most liked travel stories! The codename for v20 is Socialtubes. Now you see it? Bingo! So where are we today? Today, in this v18 major release, we have started with liking by travelers. A sequence of minor releases will follow as we will be getting more data and rolling out more parts of the product. Stay tuned and always follow your curiosity.
Bug Picture
Still, there are bugs. Old and new bugs, in data and in code. There are performance issues that may affect usability. Particularly, My Trips carousel is loading quite slowly. It’s a double-digit number of bugs. We hope you will not encounter many of them in your travel planning. Not because you are not using Curiosio but because the bugs will not pop up during your workflows.
We tested in Chrome, Chromium, DuckDuckGo browsers. The website still doesn’t work in Firefox if the search request takes 30s or longer. We have not fixed the known networking issue with the newer Firefoxes yet. You could help us with testing in Safari browsers on big and small screens. Meanwhile, we will do a series of v18 minor releases with more scope implemented and more bugs fixed.
Thank you for reading/skimming/scrolling down here. Curiosio is being built in Ukraine. If you live in a democratic and developed part of the world — you can help Ukraine to win the war against the terrorist state of Russia and their ally Belarus. Spread the word about their war crimes in Ukraine. Ask your governments to send weapons to Ukraine and to share intelligence. Ukrainians are brave and are defending well, we just don’t have the weapons against such a big atrocious enemy. Thank you in advance!
